305

Paleontology and the Recapitulation Theory.

By E. R. Cumings.

I.

In reply to a severe critique of the recapitulation theory, or biogenetic

law. by Hurst (30), Bather remarks that "If the embryologists had not forestalled them, the paleontologists would have had to invent the theory

of recapitulation." (1) This may be considered as a fair sample of the general attitude of paleontologists of the Hyatt school, to which Bather belongs, toward the recapitulation theory. Even the more conservative paleontologists, while inclined to use the

theory cum grano salts, recognize the weight of evidence that Hyatt and his coworkers in the realm of paleobiology, have brought together, as is

evidenced by the following quotation from Zittel (65) : "Nevertheless embyrcnic types are not entirely wanting among invertebrates. The Pale- ozoic Belinurida? are bewilderingly like the larvae of the living Limulus.

The pentacrinoid larva of Antedon is nearer many fossil crinoids than the full grown animal Among pelecypods the stages of early youth of oysters and Pectinidse may be compared with Paleozoic Aviculida?. Among brachiopods, according to Beecher. the stages which lhing Tere- hratulida* pass through in the development of their arm-skeleton correspond with a number of fossil genera. The beautiful researches of Hyatt, Wiir- tenburger and Branco. have shown that all Ammonites and Ceratites pass through a goniatite stage, and that the inner whorls of an Ammonite con- stantly resemble, in form, ornament and suture line the adult condition of some previously existing genus or other."

In violent contrast with this full acceptance, or this guarded ac- ceptance of the theory on the part of the paleontologists, is the position of a considerable school of embryologists and zoologists. Perhaps no one lias put the case against the theory more baldly and forcibly than Mont- gomery in his recent book on "An Analysis of Racial Descent" (42). He says: "The method is wrong in principle, to compare an adult stage of one organism with an immature stage of another." And again: "There- fore we can only conclude that the embryogeny does not furnish any re- capitulation of the phylogeny. not even a recapitulation marred at occa- [20—230031 306

sional points by secondary changes. . . . An analysis of the stages

during the life of one individual can in no way present a knowledge of its ancestry, and the method of comparing non-correspondent stages of two

species is wrong in principle." Equally sweeping is the statement of

Hurst (30) : "The ontogeny is not an epitome of the phylogeny, is not

even a modified or 'falsified' epitome, is not a record, either perfect or im-

perfect of past history, is not a recapitulation of evolution."

It would seem as though two statements could not be more flatly con- tradictory than these of Hurst and Montgomery, and that of Bather quoted

above. Nevertheless I venture to make the seemingly paradoxical asser-

tion that both parties to the controversy may be right, for the simple rea- son that they are talking about quite different things. This has been

nowhere better expressed than by Grabau (25). He says: "It has been

the general custom to test the validity of the recapitulation theory by the

embryological method; i. e.. the comparableness of the changes which the

individual undergoes during its embryonic period to the adults of more primitive types. Usually the comparison has been with the adults of ex-

isting types, since in most eases these alone were available for compari-

son. It is no wonder, then, that such comparisons have led to innumer-

able errors, if not absurdities, which have placed the recapitulation theory

in an evil light and awakened in the minds of many serious investigators

doubts as to the validity of the deductions based upon this doctrine. When,

however, the entire life history of the individual is considered, instead of only the embryonic period, and when the successive stages of epembryonic development are compared with the adult characters of related types, in immediately preceding geologic periods, it will be found that the funda- mental principle of recapitulation is sound, and that the individuals do repeat in their own epembryonic development the characters of their own immediate ancestors." (Italics mine.)

It is as a matter of fact true that the Hyatt school of paleontologists have based their phylogenies on epembryonic rather than embryonic stages —stages beginning with the nepionic or infantile—since in the nature of the case the true embryonic stages are scarcely ever accessible to the stu- dent of fossils. It is no less tine that the severest critics of the theory of recapitulation have rested their case largely en the real or supposed lack of correspondence between the embryonic stages and the adult stages of assumed ancestors, or upon certain a priori considerations having to 307 do with the laws of development and inheritance. To the former class belong such critics as Von Baer, and to the latter class such as Hatschek, His, Hurst, Montgomery and others. In making this statement I am aware that paleontologists sometimes compare true embryonic stages with adult stages of pre-existing types. As examples of this we might cite the comparison of the larval stage of Antedon with adult Paleozoic crinoids, as mentioned by Zittel; and the classic attempt of Beecher to reconstruct the ancestor of the Brachiopoda by a comparison of the phylembryonic stages of a representative series of genera of recent and fossil brachiopods. Nevertheless by far the greater number of comparisons that have been instituted by paleontologists have been between epembryonic stages of individuals and adult stages of older forms. Such comparisons are those of Hyatt, Branco, Karpinsky, Wiirten- burger, Bnckman, Neuinayr, Smith, Beecher, Clarke and others among the

Cephalopoda ; of Beecher and Schuchert, Raymond, Greene and Cumings amoug the Brachiopoda ; of Jackson among the Pelecypoda ; of Grabau and Burnett Smith among the Gastropoda; of Lang and Cumings among

the Bryozoa ; of Ruedemann among the graptolites ; and of Beecher,

Girty, Lang and others among the corals. To many of these researches I shall refer later.

I am also not unmindful of the fact that many of those who are not primarily paleontologists recognize the fact that development does not terminate with the completion of the embryonic stages, and that recapitu- lation may be legitimately looked for in epembryonic as well as embryonic

stages, or that it may be sought iu epembryonic stages, even though masked

or falsified in embryonic stages. It is true, of course, that some speak of a comparison of ontogeny and phylogeny when, judging by the context, they mean a comparison between embryogeny and phylogeny. There arises

here a question of definition : does the biogenetic law mean that the

ontogeny is a recapitulation of the phylogeny, or does it mean that the

embryogeny is a recapitulation of the phylogeny? If we take the general

consensus of opinion we shall find for the former definition, and if we

take the words of Haeckel, whose statement of the law is the one usually

quoted, we shall again find for the former definition. I believe that, as

a matter of fact, no one would maintain that the second definition is cor-

rect, however much he might forget in his studies to take the epembryonic stages into consideration. 308

Nor would I create the impression that the einbryologists and zoolo-

gists have utterly deserted the paleontologists in their support of the re- capitulation theory. Several recent papers give considerable aid and com-

fort to those of us who still believe in recapitulation. I shall introduce here the conclusions of three of these workers, more particularly because

it will afford me an opportunity to correct what I hold to be another error of those who oppose the theory.

One of the most interesting pieces of evidence that has recently been

adduced in favor of the idea that ontogeny recapitulates phylogeny is to

be found in a paper by Griggs on "Juvenile Kelps" (28). It is not my purpose, however, to discuss the very interesting evidence which he has recorded, but rather to quote his remarks on the views of His and Morgan, and his general conclusions. His maintains that the reason why ontogeny

seems to recapitulate phylogeny is because the stages in development are.

as Griggs paraphrases it, "only the physiologically necessary steps for the

formation of the adult body from its earliest stage, which in most cases

is the egg." With the ideas of Morgan as expressed in his valuable book

on "Evolution and Adaptation" we are all familiar. He holds that or- ganisms repeat in their development, not adult stages, but only embryonic stages of their ancestors. To this idea he has given the name of "repe-

tition."

On this point of the recapitulation of embryonic conditions Griggs

makes the following pertinent statements : "In the toothless animals, the

whale and the bird, the development of teeth in the jaws is entirely un-

necessary * * * it may even be said to hinder the attainment of the

adult condition. The same is true of the mammalian gill slits and of

most structures which have in the past attracted attention in connection with the recapitulation theory. As the ancestral period when such struc- tures were fully developed in the adult becomes more and more remote,

the tendency to inherit them becomes less and less, because of the cumu-

lative impulses given to the heritage by the nearer ancestors. Conse-

quently they are successively less and less developed. Any gradual' loss of inheritance can, in the nature of the case, take place only from the

mature condition backward toward the beginning of the life cycle ; other- wise we should have adult structures with no ontogenetic history. There-

fore we can understand why it is that in many cases only the embryonic

stages of ancestral history persist in the ontogeny." In a foot note he

says : "The cutting off of end stages in the development of organs has 309 given rise to the idea that the adult stages are 'pushed back into the embryo.' Such a misconception easily arose from the loose language in which the facts have often been expressed. Thus the embryogeny will be gradually shortened by the omission of more and more of the superfluous

ancestral stages ; and it will tend finally to retain only such stages as are necessary to the attainment of the adult form." Morgan and His, he maintains, have confused morphology and physiology. "The recapitula-

tion theory has nothing to do with physiology ; it is purely a matter of morphology."

In conclusion Griggs says: "Taking all the evidence into considera- tion, it seems to the writer that we are bound to conclude that though organisms are subject to adaptations at any stage of their life cycles and may gradually cut out superfluous stages, yet, except as some such ten- dency has operated to change the heritage, the development of the indi- vidual does recapitulate the history of the race * * * recapitulation must take place if there is any force which tends to make offspring like parent, if heredity is of any importance in moulding the forms of organ- isms. On the other hand, if there is any variability of transmutation of individuals in stages other than the adult end stages of the life cycles, the recapitulation cannot be perfect, but must be marred at every stage

where secondary change has taken place." I shall return later to some of the points raised by Griggs in the above statements. Another eminent worker, Dr. Eigenmann, says at the close of a paper

on the eyes of the blind vertebrates of North America (20) : "We have seen in the preceding pages that the foundations of the eye [of Amohjop-

sls'] are normally laid, but that the superstructure instead of continuing

the plan with new material, completes it out of the material provided for

the foundations, and that in fact not even all of this (lens) material en-

ters into the structure of the adult eye. The development of the founda-

tions of the eye is phylogenic, the stages beyond the foundations are di-

rect."

The third writer, Dr. Zeleny (64), in his paper on "Compensatory Regulation." in a discussion of the development and regeneration of the

opercula in serpulids. says that the morphologic series is so complete as

to make sufficient ground for the conclusion that the opercula arose in

the course of phylogeny as modified branchia. The ontogenetic series, he

says, corresponds very closely icith the prooaole pliylogenctic series.

Speaking of the regeneratory development he says : "The course of re- 310 generatory development is characterized by great condensation and direct- ness of the development. There is no trace of the branchial stage, and the development of the two rows of processes of the terminal cup does not follow the ontogenetic order."

His final conclusion is as follows : "The data furnished, therefore, by the opercula of the serpulids" give a fairly close agreement between the ontogenetic stages and the probable phylogenetic ones as determined by the usual criteria. The regeneratory development, however, follows a course which may be modified by the character of the operation that leads to the regeneration." By the "usual criteria" he means morphology, etc.. so that he cannot be accused of the circulus vitiosus.

Those who wish to review the detailed evidence given in the above papers, bearing on the theory of recapitulation, will, of course, consult the original papers. My main reason for quoting them is, as stated above, because of their bearing on what seem to me to be grave errors in the reasoning of His. Morgan and Montgomery and others who have adopted similar views. The error seems to me to be, as pointed out by Griggs, in the confusion of morphology and physiology. The. adult characters that are supposed to be recapitulated in the ontogeny, as well as the characters in ontogeny that are supposed to represent them, are morphological characters solely. It matters not what new function they may have come to serve, nor by what physiologic process they have come to make their appearance in the recapitulating organism. The confusion arising from this source colors all the argument of Montgomery, in which he endeav- ors to prove that new specific characters must have some representation in the ovum—a view which we must certainly agree with—and that there- fore "the whole row" of cells from the ovum to the adult must be differ- ent. We grant that "The whole row" is different in some way. physiolog-

ically different, different in its play of energies ; but it may conceivably be morphologically identical up to the very point where the new character is added. It is just as easy to conceive that the energy, or whatever we choose to call it, that is at a certain stage in develop- ment to produce a certain rib or spine or color-band on the shell of a gastropod, may be handed through the row of cells reaching up to the given stage, without producing a single recognizable morphologic change in the row, as compared with the individual that is not to possess the new character, as it is to conceive the opposite. The argument for the one view is just as certainly a priori as the argument for the other view. It 311

is also perfectly conceivable that the morphology of the individual cells

in the row might differ after the acquisition of the new character (in so

far as this assumption is required by recent cytological studies), and yet not a single organ or part of the organism be different up to the stage in ontogeny when the new character appears. Unless, therefore, a change in the energies of the cells inevitably necessitates a change in the morphology

of all the cells or of all the organs which they compose, the argument of Montgomery proves nothing.

As to the argument of His and others, that the supposedly ances-

tral stages are merely the physiologically necessary stages in the develop-

ment of the individual; it again, as Griggs points out, confuses mor-

phology with physiology, and is open to the further objection that it is

directly opposed to the facts. Why, for example, should the condition of

perfect blindness, with complete loss of all the essential structures of

the eye, be attainable only by the round-about way of first developing the foundations of a normal-eye? Why should a serpulid be able to regenerate a perfect operculum in a manner entirely different from, and even opposed to the ontogenesis of the organ, if there is any physiologically necessary way in which that particular individual or that particular organ must develop? The thing that makes it necessary for development to take a

certain course in a given individual is the fact that the development has

taken that same course in the ancestors. This species of coercion may,

to be sure, be relaxed, and the development take some other course, but

it is usually relaxed with extreme slowness, and after many generations have passed.

If inheritance were perfect, the individual would take exactly the same course in development as its ancestors. That it does not do this

in all cases is, as Griggs points out, a more remarkable fact than that in other cases it should follow the ancestral mode of development so closely.

Griggs explains the loss of inheritance as due to a progressive condensa- tion of the ontogeny by the "omission of more and more of the super-

fluous ancestral stages." This is the well-known law of acceleration or tachygenesis. Like most embryologists, however, he misconceives the law. as shown by the foot-note quoted above. Embryologists are especially

prone to limit the law of acceleration in development to the skipping or omission of steps, aud the consequent shortening of development. This is not in keeping with the views of Hyatt, who first definitely formulated the law; and, as all paleobiologists know, it is not in keeping with the 312

facts. Hyatt (31) says: "All modifications and variations in progressive

series tend to appear first in the adolescent or adult stages of growth,

and then to be inherited in successive descendants at earlier stages ac-

cording to the law of acceleration, until they either become embryonic or are crowded out of the organization and replaced in the development by

characters of later origin." A more concise statement of the law is as

follows : "The substages of development in ontogeny are the bearers of

distal characters in inverse proportion and of proximal characters in

direct proportion to their removal in time and position from the proto-

conch, or last embryonic stage" (31). According to the definitions just quoted, acceleration involves not

only the omission of characters, in some cases (and this is the only sort

of acceleration that most embryologists seem to recognize), but it involves also .condensation without omission, by crowding more into a given por- tion of the ontogeny, or again by "telescoping" of characters, as Grabau

(25) calls it. so that characters that originally appeared in succession, come to appear simultaneously. In other words acceleration may be by

elimination, by condensation, without change in the order of appearance of

characters, or. third, by telescoping. The latter is condensation with

change in the order of appearance, or as commonly expressed, unequal

acceleration. It is probable that paleobiologists have erred in giving too

much emphasis to the principle of earlier inheritance, involved in the

law. just as embryologists and morphologists have erred in entirely neg- lecting this phase of inheritance. As conceived by the paleobiologist, the

law of acceleration is an explanation of recapitulation, as well as an ex- planation of the failure to recapitulate.

Another factor in inheritance has been ghr en the name of retardation by Cope (15). By the operation of this factor, characters that appear

late in the ontogeny may disappear in descendants because development

terminates before the given character is reached. In this way. it is con- ceived, the ontogeny may be shortened and simplified, and many ances- tral characters lost entirely. The result of the continued operation of

retardation would be retrogression. That is. the given form, if it con-

tinued to repeat the remoter ancestral stages in the early part of its ontogeny, and continued at the same time to drop off the later ancestral stages, by failing to proceed far enough in its development, would ulti- mately come to resemble the remote rather than the nearer ancestors.

Manifestly the retarded forms do not recapitulate the lost characters, so 313

that here, also, as in the omission of characters in the earlier stages of

ontogeny, the heritage is incomplete. Of the complications of inheritance that arise from larval adapta- tions, intra-uterine adaptations, and special adaptations arising in later

life, I shall not speak. All of these have been repeatedly discussed (see

for example Smith 57), and are well understood. Against all of these the paleobiologist must be on his guard. All of these factors tend to make the parallelism between ontogeny and phylogeny inexact, as long ago

pointed out by Cope (15). Yet in spite of the operation of these factors,

the cases in which there is clear evidence of recapitulation are so numer-

ous, and so well known to the paleobiologist, that were it not for the

continually reiterated statements of certain embryologists that there is

no such thing as recapitulation, I should hesitate to again point them

out. I shall now take up the evidence according to the groups of or-

ganisms in which it has been ascertained ; and I once more remind the reader that most of this evidence applies to the epembryonic and not to the embryonic stages.

II.

Cephalopoda.—The only existing representative of the great group of Tetrabranchiata, the class to which nearly all of the fossil cephalopods

belong, is the Nautilus. The genus Nautilus is a striking example of the

persistence of a primitive type. It belongs to the more primitive branch of the tetrabranchs, from' which, according to all the evidence, the marvel- ously complex ammonities, on the one hand, and the modern naked cepha- lopods are descended. Nautilus is the only tetrabranch of Avhich the entire ontogeny, including the embryonic stages, is known.

This lack, however, in the case of the fossil genera is not as serious as might be supposed, for the reason that even in these ancient forms all of the growth stages from the latest embryonic (phylembryonic) stage to the adult are preserved in every complete individual shell. An inspection of the Nautilus shell makes this at once apparent, for the earlier stages of the shell are surrounded and protected by the later, and no part of the shell is lost or resorbed. In the straight and loosely coiled shells only, such for example as Orthoceras, Gyrtoceras, etc., is the case different; and even here, barring injury, or the dehiscence of the earlier chambers, every post-embryonic stage is preserved. From a study, therefore, of a single shell, we are able to make out perfectly all of the epembryonic de- 314 velopnient iu that part of the organism that was most vitally affected by the environment, and which must therefore indicate most perfectly the lines along which the evolution proceeded.

If the initial portion of the shell of Nautilus be examined, it will be found to be characterized by a scar or cicatrix. In the same region of the shells of ammonites and some Xautiloidea (Orthoceras) , instead of this cicatrix, there is present a minute, bulbous or bag-like shell, attached to the apex of the shell proper. If in the case of Orthoceras, as shown by

Hyatt (31), this bulb, or protoconch be broken away, there is exposed a scar (cicatrix) precisely similar to that of Nautilus. The initial shell or protoconch is therefore substantially' the same in all of the Tetrabranch- iata, and is supposed to point to a "septa-less and chamberless form simi- lar to the protoconch" as the common ancestor of these two great divisions

of the Tetrabranchiata : and possibly, as Hyatt suggests of the Cephalo- poda, Pteropoda and Gastropoda (31). The protoconch represents the latest of the true embryonic stages, namely the phylembryo.

Succeeding this early stage are the stages of the shell proper. 1 In

Nautilus the early nepionie portion of the shell, which includes the forma- tion of the first three septa, is only slightly curved ( cyrtoceraform ) . Up to the stage of the formation of the second septum, the shell is in fact nearly straight (orthorceraform). The first septum has an apically di- rected caecum, and the second septum an apically directed closed tube, the closed apical end of which fits into the caecum of the first septum.

This tube is the beginning of the siphuncle. Since the tube fits closely into the caecum, the two together form a continuous tube, in which the apical end or bottom of the siphuncular tube forms a partition or septum, so that as Hyatt points out, the resemblance "of this early stage to the adult structures of Diphragmoceras becomes perfectly clear." (31)

In the later nepionie stages (i. e., after the formation of the third septum) the shell is rather sharply bent (the gyroceran curve), so that near the close of the first volution the whorl is brought back into contact with the apex of the conch. This manner of growth results in leaving an empty space or umbilical perforation between the two halves of the first volution. In the ancient coiled Nautiloidea there appears at the be- ginning of this (neanic) stage, when the whorls come into contact, a de-

1 The stages from this point on are termed by Hyatt (31), and following him by practically all paleobiologists at tbe present time, the nepionie, neanic, epliebic and gerontic stages ; meaning respectively, the infantile, youthful, mature and old age stages of growth. 315

pression or groove in the dorsum of the whorl, where it rests against the

venter of the preceding whorl. This is the impressed zone. In the mod- ern Nautilus, however, this furrow or impressed zone begins in the early

nepionic stage, before the whorls have come into contact. This occurs

also in the nautilian shells of the Carboniferous, Jurassic, Cretaceous and Tertiary. Of this truly remarkable feature of cephalopod development, Hyatt

says : "When one ascends in the same genetic series to the more special- ized nautilian involved shells this purely acquired character becomes, through the action of tachygenesis, forced back, appearing as a rule in

the nepionic stage before the whorls touch. It is therefore, in these forms entirely independent of the mechanical cause, the pressure of one whorl

upon another, which first originated it. One need only add that this

configuration of the dorsum is never found in the adults of any ancient

and normally uncoiled shells, so far as I know, nor so far as have been figured." (31) Without reviewing any of the further interesting details of the on-

togeny of Nautilus, enough has been said to make it evident that if there

is any truth in recapitulation, the development of Nautilus would indicate

(disregarding the protoconchal characters) an ancestral line that con-

tained, first straight or slightly arched, then loosely coiled, and finally

closely coiled shells, and that the earliest of these possessed a septate siphimcle. That the geological series of shells indicates the same thing every paleontologist knows perfectly well. The development of Nautilus also affords one of the most perfect illustrations of the law of tachygene-

sis, in the earlier inheritance of the impressed zone, known in the whole animal kingdom.

One further illustration, from the Cephalopoda, of the parallelism of

ontogeny and phylogeny must suffice. This illustration is drawn from the genus Placenticeras, one of the complex Ammonites of the Cretaceous. The development of this genus has been beautifully worked out by Professor

J. P. Smith (58). The species P. pacific-it m comes from the Chico forma- tion of the Upper Cretaceous. The following account applies to the de-

velopment of this species and is drawn from the paper by Smith, cited above. The earliest shelled stage was probably passed before the animal was

hatched. This is the protoronch or phylembryo. It is a smooth, oval,

bulbous body, similar to that of all the later ammonites. It probably rep- 316

resents an "adaptive form, clue to life in the egg, and does not represent any ancient ancestral genus, for none of the early cephalopods were

shaped like this."

"With the formation of the first septum, the young ammonite has taken

its place among the chambered cephalopods, and has become, for the time

being, a nautiloid. although it is not possible .... to correlate it with any special genus The first septum .... is nau- tilian in character, but the siphuncle begins inside the protoconch with a siphonal knob, or caecum, and the protoconch itself is calcareous. These are two characters that the nautiloids even to this day, have never yet ac- quired We have in this stage ammonite characters pushed back by unequal acceleration [telescoping], until they occur contempo- raneously with more remote ancestral characters."

There is no sign of an umbilical perforation as in the Nautilus, de- scribed above, a fact which again shows the degree of acceleration of these ammonites. With the second septum the ammonite characters are assumed. The

shell at this stage is "distinctly goniatitic," but also possesses characters, introduced by acceleration, that belong to later genera. The evidence

indicating the goniatitic as well as later stages to be mentioned, is mainly

the character of the suture lines. "At about five-eighths of a coil .... the larva has reached a stage correlative with the goniatites of the Upper

Carboniferous." This stage is quickly passed, and the goniatitic char-

acters are lost and characters transitional to the ammonite stage make

their appearance. "At one and one-twelfth coils the shell is transitional from the glyphioceran stage to what resembles closely the genus X a unites

of the Trias." In regard to this stage Smith says : "If it had not been said that this was a minute shell taken out of an older individual, any

paleontologist would refer it without hesitation to the Glyphioceratidre.

and probably to ... . Pronannites, of the Lower Carboniferous." This stage lasts about one-half revolution.

In the neanic stage, at one and seven-twelfths coils, the shell re- sembles very strongly Gymbites, or some related genus of the Lower Ju-

rassic. The first signs of shell sculpture occur in this stage. In the next stage the sculpture becomes stronger, and the shell assumes a de-

cidedly aegoceran appearance. From two up to two and one-quarter coils,

the shell resembles in most respects the stock to which Perisptiinctes be-

longs, and this is accordingly called the perisphinctes stage. During this 317 stage the sides of the shell become more flattened, and the abdominal shoulders squarer, the varices frequent, and strong intermediate ribs ap- pear on the sides and abdomen. In the next (Cosmoccras) stage "the ribs no longer cross the abdo- men, but end in tubercles on the abdominal shoulders, forming well de- fined shoulder keels, with a furrow between them." Near the beginning of the fourth coil the ribs are reduced to mere faint undulations and fine sickle-shaped striae on the sides of the umbilicus, while the external tu- bercles become almost obsolete, forming mere notches on the continuous abdominal keels. Specific characters begin to appear here. This may be taken as the beginning of the Hoplites stage. The septa have not reached the complete development of the genus. The umbilical knots begin at this stage, and growing stronger, become a characteristic feature of the adult Placenticeras. "Placenticeras paclfi- cum at this stage is wholly unlike P. californicum, with which it is asso- ciated, being much more compressed and discoidal, with narrow abdo- men, flatter sides, much less distinct sculpture, and narrower umbilicus, although in the earlier adolescent periods both species are very much alike." The shell passes from this stage by gradual changes into the adult Placenticeras.

Professor Smith's conclusions are of especial interest. He says : "The development of Placenticeras shows that it is possible, in spite of dog- matic assertions to the contrary, to decipher the race history of an animal in its individual ontogeny." 1

1 For further illustrations of recapitulation among the Cephalopoda, the stu- dent should consult tbe following papers : Branco, W., Beitrage zur Entwicklungs- geschichte der fossilen Cephalopoden, Paleontographica, vols, xxvi, xxvii, 1879, '80. Buckman, S. S., Monograph of the Inferior Oolite Ammonites, Paleontographical Society, 1887-'96. Hyatt, A., Paralellism of the individual and the order among tetrabranchiate Moliusks, Mem. Bos. Sac. Nat. Hist., vol. i, 1866 ; Fossil cephalo- pods of the Museum of Comparative Zoology, Bull. Mus. Zool., vol. iii, 1872 Comp. ;

Genesis of the Arietidse, Smithsonian Gontr. to Knowl., vol. xxvi, 1889 ; Phylogeny of an acquired characteristic, 1'roc. Am. Phil. Soc, vol. xxxii ; Cephalopoda, in Text Boole of Paleontology by Zittel (Eastman trans.), 1899. Hyatt, A., and Smith, J. P., Triassic cephalopod genera of North America, U. S. Geol. Surv. Prof. Paper No. 40, 1905. Karpinsky, A., Ueber die Ammoneien der Artinsk-Stufe, Mem. Acad. Set. Imp. St. Petersburg, vol. xxxvii, No. 2, 1889. Neumayr, M., Die Ammoniten der Kreide und die Systematik der Ammonitiden, Zeitschr. der Dcuich. Geol. Ges.,

1875 ; Ueber unvermittelt auftretende Cephalopodentypen im Jura Mittel-Europas,

Jahrb. d. K. K. Geol. Reichs. Wien, vol. xxviii, 1878. Smith, .T. P., The development of Glyphioceras and the phylogeny of the Glyphioceratidse, Proc. Calif. Acad. ScL,

(3) Geol., vol. i, 1897; The Development of Lytoceras and Phylloceras, Ibid..

1898 : Larval stages of Schloenbachia, Jour. Morphology, vol. xvi. 1899 ; The Car- boniferous Ammonoids of America, Monog. TJ. S. G. S.. No. xlii. 1903. Wiirtenburger, R., Studien iiber die Stammgeschichte der Ammoniten, Leipzic, 1880. 318

Pelecypocla. —The classic memoir of Jackson (32) on the phylogeny of the Pelecypocla brings together numerous illustrations of recapitula- tion among the members of this class of animals. Jackson's conclusions are well-known, and I shall therefore review them very briefly. From a study of a large number of genera representing widely diver- gent members of the Pelecypoda, Jackson concludes that there is present throughout the group an embryonic shell, which he calls the "prodisso- conch" (a term correlative with the term protoconch of the Cephalopoda and Gastropoda ),, and which is a simple bivalved, equivalve shell. At this (phylembryoriic) stage of development there are two adductor muscles, even in genera in which the adult have only one adductor. That is, the prodissoconch is dimyarian even though the adult animal may be mono- myarian. In the Aviculidre and their allies (Ostrea, Avicula, Perna, Pec- tvn, Plicatula, Anomia) the prodissoconch very closely resembles in form the primitive genus Nucula. The anatomical characters of the prodisso- conch also bear out this resemblance. It is therefore inferred that some such type as Nucula is the primitive ancestor of the Aviculida?, and pos- sibly of the Pelecypoda. The paleontological and anatomical evidence supports this conclusion.

We have here, then, in the Aviculidre and their allies, a group of monomyarians, some of them, as Ostrea, Plicatula, and Anomia, of very aberrant form, the representation in the ontogeny of a dimyarian stage, which, from all the evidence, actually characterized the adults of the ancestors of the group. Whether or not Nucula is the actual ancestor of this group of pelecypoda, it is quite certain that the earliest pelecypods were of the same general form as the prodissoconch, and that they were dimyarian.

In the same paper Jackson has shown in a masterly manner that the ostreaform shape of the shell, which characterizes many more or less widely separated genera of pelecypods, is due to "the mechanical con- ditions of direct cemented fixation." These ostreaform shells are very variously derived, and should, if there is anything in the theory Of re- capitulation, each show in the young stages, before the valves have be- come fixed, the distinctive adult characters of its particular ancestor. In this case we are relieved from the danger of arguing in a circle by the fact that the genetic relations of most of the forms are fairly well known from lines of evidence other than the ontogeny. The following specific cases cited by Jackson are of especial interest. 319

Mulleria lohata, a member of the Unionidse, "is so remarkably like an

oyster (in the adult) that it has been called the fresh-water oyster. In

the monomyarian adult .... the shell is rough and irregular with a dee]> attached and flatfish free valve, and a specimen in the Museum of

Comparative Zoology is indistinguishable in shape from forms commonly

found in Ostrea virginiana The young shell of Mulleria

. . . . is Anodon-shaped, equivalvular and dimyarian as described by authors."

Hinnites is another genus which has the ostreaform adult. "In the

young it is free and peetiniform, but in the adult .... so close

is the likeness to an oyster that in the synonomy of the genus it has been named Ostrea and Ostracites." In Hinnites cortesi of the Tertiary, in

the neanic stage, the right valve is purely peetiniform. "It has the well-

developed ears, deep bj^ssal sinus, and an evenly plicated shell which at

this stage is nearly or quite equivalvular." With the period of attach-

ment a most marked change in the valves takes place and the adult be- comes deeply concave (in the right attached valve) and highly ostrea-

form. The byssal notch is filled up and "completely wiped out of exist- ence."

In genera such as Ostrea and Plicatula, where fixation takes place at the close of the prodissoconch stage, the succeeding stages give very

little indication of the ancestry, owing to the extensive modification of the

shell as soon as fixation takes place. According to Dall Ostrea is derived

from tlie Pteriidae.

Spondylus is another genus in which cementation has caused exten- sive modification of the valves in the adult. Fixation takes place at the close of the nepionic period. Therefore this genus may be expected to

afford some evidence of recapitulation. The first nepionic stage of Spon-

dylus is decidedly peetiniform. It has a long hinge-line and a deep byssal sinus. After fixation, in the first stages of irregular growth, the byssal notch is soldered over, and eradicated in a manner similar to Hinnites.

Another illustration of recapitulation among the Pelecypoda is the

case of Pecten itself. Of this genus Jackson says : "In the development

of the modern Pecten we find in the first stages of dissoconch growth a form of shell .... presenting characters which make it referable in ancestral origin to Rhomftopteria, a member of the true Aviculidte, later succeeded by a growth .... bearing marked features referable in origin to an ancestral genus Pterinopecten Still later a stage 320 exists .... which is referable in its inherited form to Aviculopec- ten, and finally the true Pectcn features characteristic of the adult .... are established. The geological sequence of these several groups is in the order indicated by the development of Pectcn. We have, therefore, a clear case of the ontogeny of an individual illustrating the phylogeny of the group." Gastropoda.—For studies of the Gastropoda in which growth stages have especially been taken into consideration we are indebted chiefly to

Grabau (22. 23, 24. 25) and Burnett Smith (53, 54. 55, 56). My illustra- tions of recapitulation among the members of this class will be drawn, therefore, from the writings of these two authors.

It is eoininonly known that the apical whorl of the gastropod shell may differ materially from the succeeding portions of the shell (conch), being smooth and without ornament in cases where the conch is highly sculptured. or in some forms, as Acmaea and Crepidula, being coiled, al- though the adult shell is patelliform and non-coiled. To this apical whorl the name "protoconch" has come to be applied, a nauie which, as we have already seen, is also applied to the embryonic shell of the Cephalopoda.

Grabau (22) has suggested the use of the name "protorteconch"' in place of protoconch for the initial shell of the gastropods.

The protoconch of the existiug Gastropoda is more variable than that of the Cephalopoda, as would be expected from the highly specialized nature of most of the extant representatives of the class. In most cases there is no definite line of demarkation between the protoconch and the conch, but in a few cases, as in Fusils, etc.. the "end of the protoconch is strongly marked by the existence of a pronounced varix and an abrupt change of ornamentation." (22) "The early whorls of the protoconch

. . . . are smooth rounded coils of the type found in adult Xaticu.

. . . . In the majority of cases the initial whorl is minute, while the succeeding ones enlarge gradually aud regularly. In some types the in- itial whorl is large and swollen This type of protoconch has been termed 'bulbous' by Dall (19). The naticoid form of protoconch is in general umbilicated, and it is probable that at least the earlier por- tion of the protoconch is umbilicated in the majority of gastropods.

"From the characters of the initial whorls of the protoconch we may argue that the radicle of the coiled gatropods must have been a naticoid type with a well-marked umbilicus. Such a type is found in Straparollina reraota Billings, one of the earliest coiled gastropods of the Etcheminian 321

or Lower Cambrian of the Atlantic border province of North America.

That it is not the most primitive type of gastropod is suggested by the con-

sideration that the earliest stage .... of the protoconch is not coiled, but rather cap-shaped like modern Patella. Such primitive types

are found in Lower Cambrian species which have variously been referred

to Platyceras, Scenella, or Stenotheca, owing to the want of sufficient

characteristics to define their exact relations." (22.)

From the above it appears that the early protoconch stages indicate an ancestor of the simple, smooth shelled, umbilicated type exemplified by

StraparolUna, and that this is actually the only type of coiled gastropod

characteristic of the basal Cambrian. It is also likely from paleontolog-

ical evidence that the very earliest type of gastropod possessed a conical

1 or cornucopia-shaped shell of the Scenella type. Such an ancestry is, according to Grabau, suggested by the cap-shaped earliest stage of the protoconch. 2 One of the most completely worked-out cases of recapitulation among Gastropoda that has come to my knowledge is that of the races of Athleta petrosa Con. and its allies. The phylogeny of this group of gastropods has been very fully studied by Burnett Smith (54), from whose paper

the following account is drawn.

1 Sardeson (50) suggests that the gastropod ancestor was an "asymmetrical long conical shell" of the pteropod type. He may be right, but even so, I do not see that his conclusion would in the least invalidate the conclusions of Grabau in regard to the phylogenetic significance of the protoconch, although Sardeson seems to think so. Grabau says very plainly that the coiled shell is probably not the most primitive type of shell, and he points out the fast (quoted above) that the initial portion of the protoconch is cap-shaped and may indicate some such remote ancestor as the Cambrian forms referred to the genera Platyceras, Stenotheca, and Scenella. Whether this patelliform ancestor was in turn derived from a long conical shell, or whether on the other hand the coiled type of shell was derived directly from the "long conical" shell without the mediation of a patelliform ancestor, does not materially affect the conclusions that at a very remote time a coiled gastropod radicle was established from which practically all modern gastropods were derived. To my mind the conclusion that the ultimate ancestor of the Gastropoda was a "long conical" shell is by no means established. 2 Burnett Smith (55) concludes from a study of the Tertiary species of the genus Athleta that "we can say for this restricted normal group at least that the apes is not only a variable feature, but the most variable feature which the shells furnish." In a footnote he says "The author is thoroughly convinced that the features of the apex must be used in classification with great caution." The varia- tions which he cites in this and other papers (54, 55, 56) seem to be chiefly in the size of the protoconch, and the degree to which acceleration has caused conchal characters to appear in the later protoconchal stages. His caution, however, in regard to the classificatory value of the protoconch, should put students of the gastropods on their guard against a too free use of this portion of the shell in the establishing of genera. [21—23003] 322

The species under consideration occur in the Gulf Eocene, extending

nearly throughout it. They have heretofore been referred to the genus

Volutilithes, but are placed by Smith (55) in the genus AthJeta. Smith states that the material at his disposal was very complete, and enabled

him to study large series of individuals, very carefully collected with ref- erence to horizon. The stratigraphy of the formations from which they

came is also well understood. These favorable conditions of study, it may be remarked, are especially important in the present connection, because they enabled Smith to trace out the evolution of the forms practically continuously from zone to zone, without being chiefly dependent on onto- geny for indications of their relationships. Another fortunate circum-

stance is the fact that this author is disposed to use the evidence from ontogeny with the utmost discretion, everywhere checking it by an ap- peal to the morphological and geological series.

In the forms under consideration, the first two or three whorls are smooth and rounded, constituting the smooth or protoconchal stage. "The first ornamental feature to appear on the smooth, rounded whorl is the transverse rib, that is, a slight elevation of the whorl which runs across it

from suture to suture. These early ribs are invariably curved slightly, and each one is simple and uniform from suture to suture. The curved ribs persist as a rule for about a quarter or a half of a whorl, or even for a much less space The curved rib stage .... has been found in every species and race dealt with in this paper. The curved ribs, after about one-third of a whorl, change abruptly into the straight ribs of what has been designated the cancellated stage."

"The cancellated condition is found more or less well developed in all the different races. In the primitive races it may persist as a con- stant feature to the end of the individual's life; but in most forms it covers only a few whorls and is more variable than the preceding curved rib stage." The end of the cancellated stage is much less definite than

the beginning. It is followed by the "spiny stage." In this stage the shoulder tubercle is sharp and spine-like. Other tubercles have disap- peared, and this portion of the shell is therefore no longer cancellated. Succeeding the spiny stage, there may be a senile stage. In the base of the Eocene at Matthew's Landing, Alabama, occurs a species, Athleta limopsis, which from its primitive characters, and its po-

sition at the base of the Eocene. Smith regards as the ancestor of the races and species which he deals with in his paper. This form presents : :

323

no stages later than the cancellated stage. There is also very little in- dividual variation. Associated with A. limopsis is the species A. rugatus.

In its earlier stages this species very closely resembles A. limopsis, but "differs radically .... from that form with the progress of its ontogeny." In its later whorls it presents evidence, though not extreme, of senility. It has no spiny stage. The next species A. petrosa, represents an assemblage of races con- nected by many intergrading forms. These races range upward from the Nanfalia beds to the Jackson beds of the Eocene. Several of them are senile races, and in the adult strikingly different from the ancestral form, A. limopsis. Smith says, however, that the young of all the races "are remarkably uniform and constant. The early whorls indicate clearly that they are all descended from a cancelated ancestor, and bear a strong re- semblance .... to the characters of A. limopsis." Some of the senile races of petrosa are profoundly modified in the adult, as for ex- ample, the Hatchetigbee race, derived from the main stock through the Bell's Landing and Wood"s Bluff races. Yet their derivation from the

main stock is shown by intermediate forms, and the young of the terminal

races greatly resemble the ancestral form. In the Jackson race, which is the terminal member of the main stock, the last two whorls are spiny, and the last whorl shows some senile characters at its close. "This race shows a regular and even ontogeny." Acceleration has carried the curved rib stage back to the beginning of the third whorl, whereas in the an- cestral A. limopsis this stage begins near the close of the fourth whorl. Smith has graphically expressed the main developmental and phylo- geuetic changes in the following diagram

' 1 2 3 4 5 6 7 8 9 1 10 11 12

i

A. limopsis.

| '" ' A B C '1" '

A. petrosa. D D A B ... c and and E Slight I I

A. petrosa. D A B C D and E Upper Eocene.

In the above diagram the figures across the top stand for the number

of the whorl of the shell, and the letters indicate the different ontogenetic stages as follows 324

A—Smooth stage. B—Curved rib stage. C—Cancellated stage. D—Spiny stage. E— Senile stage.

I-—-Individual variation.

The acceleration of the Jackson race is beautifully brought out in

this diagram, and as its correlative, the recapitulation in the earlier onto- geny of the later races, of the adult characters of the ancestral race. The

individual variations may occur on any part of the shell, but usually fol-

low stage C. 1

Bracliiopoda.—Among the members of this class there is a wealth of

illustrations of recapitulation. I can only select a few cases that have

been worked out in such a way that the relationships of the forms are in- dicated by the morphological and geological series as well as by the on- togeny. The pioneer student of the correlation of ontogeny and phylogeny among the brachiopods was Beecher, whose refined researches in paleo- biology have never been excelled and rarely equaled. The developing brachiopod, in the later embryonic stages, secretes in the mantle on opposite sides of the body two shell plates, which by periph- eral growth ultimately meet at the edges and form the initial shelly

investment of the animal. This initial shell to which Beecher has given the name "protegulum" (6) is of very simple form, consisting substan- tially of two convex plates of semicircular plan, gaping at the posterior straight edges. Through this gap between the two valves the pedicle

(organ of attachment) projects. At first the pedicle occupies the full width of the valves, but subsequent peripheral growth of the shell with-

1 For additional studies of the gastropoda from the developmental standpoint see the following: Koken, E., Ueber der Gastropoden voni Cambrinm bis zur Trias.. Jahrh. fiir Mineral. G-eol. u. Pal., 1SS9. Beil. Bd. vi. Linden, Grafen M. von. Die Entwicklung der Skulptur und der Zeichnung bei den Gehausschnecken des Meeres, Zeitschr. Wiss. Zool., vol. Ixi. 1896. Grabau. A. W., Studies of Gastro- poda II, Fulgar and Sycotypus, Am. Nat., vol. xxxvii, 1903 ; Phylogeny of Fusus and its allies, Smithsonian Miscell. Coll., vol. xliv, 1904 : Studies of Gastropoda III on Orthogenetic variation, Am. Nat., vol. xli, 1907. Smith, Burnett, Phylogeny of the species of Fulgur with remarks on an abnormal specimen of Fulgur canal- iculatum and sexual dimorphism in Fulgur carica, Proc. Acad. Nat. Sci. Phila., vol. liv, 1902; Senility among Gastropods, Proc. Acad. Xat. Sci. Phila., vol. lvii, 1905: Phylogeny of the races of Volutilithes petrosus, Proc. Acad. Sat. Sci. Phila., March, petrosa, 1906 ; A new species of Athleta and a note on the morphology of Athleta Proc. Acad. Xat. Sci. Phila.. May. 1907; A contribution to the morphology of Pyrula, Proc. Acad. Sat. Sci. Phila., May, 1907. 325

out corresponding enlargement of the pedicle, leaves the latter restricted to a notch (delthyrium) in the posterior margins of the valves, providing the peripheral growth is about equal on all anterior and lateral radii.

If the shell growth is greater in the anterior direction, the shell becomes pointed, the pedicle (posterior) end remaining of about the original width. If the shell growth is mainly in the lateral directions, the shell becomes wide, with a long straight hinge, of which the pedicle opening forms a very small proportion. Whatever may be the later growth of the shell, all the earlier stages are preserved, except in cases where the beaks are injured or resorbed by the encroachment of the pedicle in adult and senile stages. The growth of the shell is entirely by additions at the mar- gins or on the inner surface. It follows that the protegulum may in ex- ceptionally well preserved material be seen intact at the beaks of the adult shell. It is often seen at the apices of young shells. Searching for the phylogenetic significance of the protegulum, Beecher

(G) ascertained that certain of the earliest known brachiopods approxi- mate very closely in form to the protegulum, and he selected the genus

Paterina (Iphidea) as the radicle of the class. It has since been shown that Paterina is not the most primitive known brachiopod. 1 It is still true, however, that the most primitive brachiopods known are of the same general form and type as Paterina, in fact they approximate more closely, if anything, than that genus, to the form of the protegulum. It may be very safely concluded, therefore, from the geological evidence, that the primitive brachiopod was actually of the type indicated by the protegulum.

Beecher says of Paterina: "In mature specimens, all lines of growth, from the nucleal shell to the margin, are unvaryingly parallel and con- centric, terminating abruptly at the cardinal line. In other words, no changes occur in the outlines or proportions of the shell during growth, through the nepionic and neanic stages up to and including the com- pleted ephebic condition. The resemblance of this form to the protegulum of other brachiopods is very marked and significant, as it represents a mature type having only the common embryonal features of other genera." Among the Brachiopoda, as among the Pelecypoda there are a number of forms in which the condition of very close fixation or of burrowing has

1 Walcott (62) seems to reserve this distinction for his genus Rustella. Pater- ina is by him made a subgenus of the genus Micromitra. These forms are all placed in the superfamily Rustellacea. 326

given rise to extremely aberrant types- One of the most extreme of these

types is the genus Probos&della. The adults of this genus bear a very

marked resemblance to the Peleeypod genus Aspergillum. In the early neanic stages Proboscidella resembles an ordinary Productus, from which

genus the type is known to have descended. Orbiculoidea is a genus

originating in the Ordovician. and extending through the Mesozoic. The

first stage is paterina-like. the second resembles Obolella, the third is like

ScJiizocrania, and adult growth brings in the characters of Orbiculoidea.

The geological order of these genera is the same as the ontogenetic order of Orbiculoidea.

Of Orbiculoidea and its allies Beecher (7) says: "The early stages of Paleozoic Orbiculoidea have straight hinge-lines and marginal beaks, and in the adult stages of the shell the beaks are usually subcentral and the growth holoperipheral. This adult diseinoid form, which originated and was acquired, through the conditions of fixation of the animals, has

been accelerated in the recent Disci nisca so that it appears in a free-swim-

ming larval stage. Thus a character acquired in adolescent and adult

stages in a Paleozoic species, through the mechanical conditions of growth,

appears by acceleration in the larval stages of later forms before the as-

sumption of the condition of fixation which first produced this character.'" In the higher genera of the Terebratellidre, the ontogeny recapitulates

the phylogeny with remarkable fidelity, as pointed out by Beecher (7).

This example has become classic, so that it is scarcely necessary to re-

peat the details. I shall give Beecher*s conclusions in his own words.

He says : "In each line of progression [the austral and boreal subfami-

lies J in the Terebratellida?. the acceleration of the period of reproduction,

by the influence of environment, threw off genera which do not go through the complete series of metamorphoses, but are otherwise fully adult and

even may show reversional tendencies due to old age : so that nearly every stage passed through by the higher genera has a fixed representa- tive in a lower genus. Moreover the lower genera are not merely equiva-

lent to or in exact parallelism with, the early stages of the higher, but they express a permanent type of structure, as far as these genera are

concerned, and after reaching maturity do not show a tendency to at- tain higher phases of development, but thicken the shell and cardinal

process, absorb the deltidial plates, and exhibit all the evidences of senility." 327

Raymond (46) has pointed out a number of interesting cases of reca- pitulation. The very common and well-known Devonian Spirifer, 8. mu- cronatus, has the cardinal extremities in the adult very acute (mucro- uate), sometimes, indeed, drawn out into needle-like points; while the number of plications may be thirty or more. In the neanic stage these transversely elongated spirifers pass through forms corresponding to the adults of certain Niagara species. The adult of S. crispus, corresponds very closely in shape, number of plications, and shell index with these young specimens of S. mucronatus.

Shinier and Grabau (51) have shown that in the upper part of the Hamilton series of Thedford, Ontario, there occurs a variety of Spirifer mucronatus, which though not mucronate at all in the adult, is "extremely mucronate" in the neanic stage. At this stage also there is evidence of the median plication of the sinus, another charactertistic of the adult of the normal 8. mucronatus. In the adult of the Thedford variety this median plication has disappeared. The geological and morphological evi- dence of the derivation of this form of 8. mucronatus is complete.

I have pointed out an exactly similar case in the variety senex of

PlatystropMa acutilirata (16). This variety occurs in the upper part of the Whitewater division of the Richmond series of Indiana and Ohio.

PlatystropMa acutilirata, as is well known, is very mucronate in the adult, resembling in its general outline, Spirifer mucronatus. It was in fact at first referred to the genus Delthyris (Spirifer). The normal form is shown by an unusually closely graded series of intermediate forms to be descended from P. laticosta, and it repeats the adult characters of the latter very faithfully in its late neanic stage, becoming always more mu- cronate as development proceeds. The upper Whitewater form, var. senex, frequently has entirely lost, in the adult stages, the acute angula- tion of the cardinal extremities, so that the lateral and cardinal edges make a right, or nearly a right angle. In the young (neanic) stages of

P. senex, however, the shell is decidedly mucronate, so that these young shells exactly resemble the normal PlatystropMa acutilirata of the lower

Whitewater and Liberty formations. P. senex, it may be remarked, is a well denned form, and its derivation from P. acutilirata is beyond ques- tion, since it is connected with the latter by every gradation. Another interesting case of recapitulation among the brachiopods has been worked out with great care by Mr. F. C. Greene (27). In this case also no pains was spared to ascertain the relationships of the various 328

forms by tracing theui continuously from zone to zone, and by a comparison of the morphological characters of the adults. The group studied by

Greene is that of Ghonetes grain/lifer, from the Upper Carboniferous rocks

of Kansas. Here the forms from the higher zones repeat in their onto-

geny the characters of forms from the lower zones with great fidelity. The very young stages also recall very forcibly the species of Ghonetes from

the Devonian. Ghonetes granulifer is also very interesting from the fact

that the first hinge-spines appear very much earlier in the ontogeny than

is the case in the Devonian species studied by Raymond (46). therefore showing a considerable degree of acceleration of this character during

the interval from the Devonian to the Upper "Carboniferous. Other interesting cases of recapitulation among brachiopods have

been pointed out by Beecher and Schuchert (12) in the development of the brachial apparatus in DieJasma and Zygospira.1 Trilooita.— Studies of the early stages of the development of trilo-

bites have been published by Barrande (3, 4), Walcott (59, 60, 61), Beecher

(8, 9). Matthew (39. 40. 41) and others, but for indication of the corre- lation of the ontogeny and the phylogeny in this class we are almost en-

tirely indebted to Beecher. In his papers oh "Larval Stages of Trilo-

bites" (8), and a "Xatural Classification of the Trilobites" (9), he has not only pointed out the remarkable way in which characters are re-

capitulated in this class, but has also proposed what is probably to be re-

garded as the most perfect example of a phylogenetic classification of a group of organisms, in existence. The earliest developmental stage of trilobites that has ever been found

(barring supposed trilobite eggs) is the larval stage or "protaspis," as it is called by Beecher (8). The protaspis is a minute body of ovate or dis- coid shape, and about a millimeter in length. This larval stage has

1 For additional examples of recapitulation among the brachiopods see the following: Beecher, C. E., Studies in Evolution (a series of collected papers), Scribners, 1901. Beecher, C. E., and Clarke, J. M., The Development of some Silurian Brachiopoda, Mem. N. T. State Mus., No. I, 1889. Beecher, C. E.', and Schuchert, C, Development of the shell and brachial supports in Dielasma and Zygospira, Proc. Biol. Soc. Washington, vol. viii, 1S93. Cumings, E. R., The

morphogenesis of Platystrophia ; A study of the Evolution of a Paleozoic Brach- iopod. Am. Jour. Sci., vol. xv, 1903. Raymond. P. E.. The developmental change

i'n some common Devonian brachiopods. Am. Jour. Sci., vol. xvii, 1904. Greene, F. C, The development of the Carboniferous brachiopod Chonetes granulifer, Owen. Jour. Geol., vol. xvi, 1908. Buckman, S. S., Homeomorphy among Jurassic Brachiopoda. Proc. Gottesicold Nat. Field Club, vol. xii, 1901. 329

been seen in a sufficiently representative series of genera to make it rea-

sonably certain that it is the common larval type among the trilobites.

It is pretty well established that the eye of crustaceans has migrated

from the ventral to the dorsal surface of the cephalon. At an interme- diate stage in this process the eyes would appear on the margins of the

cephalon. If this has been the history of the eye. the most primitive larvae should show no evidence of eyes on the dorsal surface, and since

the eye is on the inner margin of the free cheek, there should be no evi-

dence of the free cheek. This is exactly the case in the youngest larvae of Ptychoparia, Solenopleura and Mostracus, "which are the most primi-

tive genera whose protaspis is known. The eye-line is present in the

later larval and adolescent stages of these genera, and persists to the

adult condition. In Sao it has been pushed forward to the earliest protas-

pis, and is also found in the two known larval stages of Triarthrus. Sao

retains the eye-line throughout life, but in Triarthrus the adult has no

traces of it, arid none of the higher and later genera studied has an eye-

line at any stage of development." This character according to Matthews,

is characteristic of the Cambrian trilobites. In its phylogenesis in later

trilobites it disappears first from the adult stages, and is finally lost from the entire ontogeny. The eyes appear on the margin of the cephalon in the last larval stage of Ptychoparia, Solenopleura, Liostracus, Sao, and

Triarthrus. In the later genera the eyes are present "in all the protaspis

stages, and persist to the mature, or ephebic condition, moving in from

the margin to near the sides of the glabella."

According to Beecher (8) "A number of genera present adult char-

acters which agree closely with some of the larval features [of later genera]. The main features of the cephalon in the simple protaspis forms

of Solenopleura, Liostracus. and Ptychoparia are retained to maturity in

such genera as Carausia and Aeon then s. which have the glabella expanded in front, joining and forming the anterior margin. They are also without

eyes or eye-line. Ctenocephalus retains the archaic glabella to maturity, and likewise shows eye-lines and the beginnings of the free cheeks (larval

Sao). Conocoryphe and Ptychoparia are still further advanced in having

the glabella rounded in front, and terminated within the margin (larva of Triarthrus). These facts and others of a similar nature show that there

are characters appearing in the adults of later and higher genera, which

successively make their appearance in the protaspis stage, sometimes to

the exclusion or modification of structures present in the most primitive 330

larvae. Thus the larvae of Dalmanites and Proetus, with their prominent eyes, and glabella distinctly terminated and rounded in front, have char- acters which do not appear in the larval stages of ancient genera, but which inay appear in their adult stages. Evidently such modifications have been acquired by the action of the law of earlier inheritance or tachygenesis."

Bryozoa.—My studies (17, 18) were the first to show that there is in the bryozoan colony a definite recapitulation of ancestral characters, and that in this particular the colony behaves as an individual. This same

fact was very clearly pointed out by Ruedemann (47) two years earlier

in the Graptolites, and I take pleasure in quoting his very explicit state- ment. He says : "Furthermore the fact that the thecal within the same

colony show a gradation from phylogenetically older to younger forms, and therefore analogous to the organ of a growing individual, pass

through ancestral stages, as, e. g., do the septa of a cephalopod shell, demonstrates how closely the zooids of this colony were united into one organism, and that practically they were more the organs of an individ-

ual than the component of a colony If the graptolites so

closely approached the morphologic value of an individual, it may be ex-

pected that, like an individual, the whole colony has its ontogeny and re- passed ancestral stages." My studies, referred to above, brought out the fact that the bryozoan colony begins as a minute hemispherical body, the "protceciuni" which is the earliest exoskeletal stage of the first individual of the colony. This protceciuni (basal disc) is very conspicuous in the Cyclostomata. and also in the ancient Cryptostornata (as shown in Fenestelhi). 1 It can not be definitely asserted that the protoechim corresponds to any ancestral bryo- zoan, but the marked resemblance of the zooeeia of some of the ancient

Stomatopora of the Ordovician to the protceciuni is at least very sug- gestive.

The ancestrula, or first complete individual of the colony, has long been known to present characters more similar to those of ancestral forms

1 I first used the term protceciuni as the designation of the first individual of the colony, and in this sense it would be exactly equivalent to the term ancestrula of Jullien. In a later paper (18) I restricted the term to the basal disc (of Barrois) which is the calicified wall of the metamorphosed and histolyzed embryo in its earliest sedentary stage. Out of this basal disc the first normal individual arises by a process stirctly analogous to budding. In this sense, therefore, the term protcecium is exactly correlative with the terms protegulum, protoconch, prodissoconch, etc. 331

than the characters of the ephebastic zooecia (see Nitsche 44, aucl Pergens

45). I have succeeded in finding evidence (18) that this is true to a no- table extent in the ancient FenesteUa. where the tubular ancestrula bears a striking resemblance to the simple tubular ephebastic zooecia of the

Cyclostomata, from which group there is every reason to believe the Cryptostomata are descended.

It is also pointed out by Nitsche and Pergens (loc. cit.) that the earlier budding habit of the colony is similar to ancestral types. In my own studies I was able to show that the early budding habit is very uni- form in the most diverse types of Bryozoa, and that it corresponds to the

budding habit that prevails throughout the astogeny of the reptant sto- matoporas. In FenesteUa my studies indicate that the earlier individuals (nepi- astic) of the colony are very different from the adult (ephebastic) indi- viduals and are strikingly similar to the ephebastic individuals of certain Cyclostomata that are on morphological grounds, as pointed out by Ulrich

(63), probably ancestral. And again, the early neanastic zooecia of the Devonian fenestellas studied are almost exactly like to the ephebastic zooecia of the fenestellas of the Niagara series. Unpublished studies in- dicate that in the Fenestellas of the Upper Carboniferous the neanastic

stage is more abbreviated, and that the adult type of zooecia follows more closely upon the nepionic type. Dr. Lang of the British Museum has published very interesting studies of the Stomatoporas and Eleids of the Mesozoic (35, 36, 37), and has come independently to exactly the same conclusions as the writer in regard to the development of the colony, and the relations of astogeny and phylo- geny among the Bryozoa. He says (35), "The development of the colony

is comparable with and follows the same laws as the development of the

individual." And again : "Among Jurassic forms of Stomatopora and

Proboscina it has been found that when any given character, such, for

instance, as the ratio of the length of the zooecium to its breadth, is fol-

lowed from the first zooecium to the last, that it has a progressive develop- ment, or anagenesis, reaches a maximum, or acme, and often may be seen to have a retrogressive development, or katagenesis, in the ultimate branches of the zoarium." Lang has paid especial attention to the manner of branching in Juras- sic stomatoporas. The nearly universal method of branching in the Juras-

sic members of this group is by dichotomy. This according to Lang may 332

be by one or other of three types as folloAvs : In type I the two zoceeia are separate throughout their entire length, only touching at their bases. In type II they are contiguous throughout their length, and in the interme- diate type they are contiguous for part of their length. To a large extent correlated with these types of dichotomy is the angle of divergence of the branches.

In all the Jurassic stomatoporas and in a few proboscinas the first dichotomy is according to t3qje I, and at a very wide angle (1S0°). The second dichotomy, in the majority of cases, is also according to type I, with an angle of 120°. The next is commonly only 90°, the next 60°, and the next 45°. all according to type I. "In primitive [Jurassic] forms the branching never gets beyond type I with a small angle. In the majority of forms, however, sooner or later the intermediate type of branching comes in. and in a great many forms this type is the final one. In a few cases of

Stomatopora. and in all Probosciiia, type II is at some time or other reached, and remains the ultimate form of branching of the zoarium. This sequence namely. Type I —Intermediate type—Type II. is invariably followed." (35).

In primitive Proboscina (a genus derived from Stomatopora) the first dichotomies are according to type 1. "In the typical forms of Proboscina the early stages have been so condensed according to the law of accelera- tion (Tachygenesis). that the first dichotomy is formed on type II.

. . . . In the more advanced types of Proboscina .... the ar- rangement of peristomes is irregular from the first." This is the typical arrangement for Bernicea, a derived genus of which Stomatopora and

Proboscina are the first two terms. It is worthy of notice that while in the Jurassic forms of Stomatopora type II is not very common, it is ex- tremely common in the Cretaceous forms. 1 Graptolites.—The beautiful researches of Ruedemann in this group have shown us. as pointed out above, that the graptolite colony closely

approaches the morphologic value of an individual, and that, like the in- dividual, it presents definite ontogenetic (astogenetic) stages. Ruedemann

C47) applies to the colonial development the terminology proposed by

1 For studies in the zoarial development of Bryozoa see Curnings, E. R.. The development of some Paleozoic Bryozoa, Am. Jour. Sci., vol. xvii. 1904 ; Develop- ment of Fenestella, Am. Jour. Sci., vol xx, 1905. Lang, W. D.. The Jurassic forms of the 'genera' Stomatopora and Proboscina, Geo}. Mag., Dec. v, vol. i, 1904 ;

The Reptant Eleid Polyzoa, Geo]. Mag. Dec. v, vol. Hi, 1906 ; Stomatopora antiqua.

Haime. and its related Liassic forms, Geol. Mag., Dec. v. vol. ii. 1905. 333

Hyatt (31). In a later paper, however, he approves the terminology in- troduced, by me, and proposes to call the development of the colony the astogeny (48).

The embryonic stage of the graptolites is represented by the initial portion of the sicula (first zooid), according to Ruedemann; and Holm

(29) asserts that the more pointed end of the sicula "corresponds to the original chitinous covering of the free zooid germ or embryo." This in- itial part of the sicula, according to Ruedemann, holds a position similar to the protoconch of the cephalopod shell.

In part I of his splendid monograph of the Graptolites (48) of New

York, at page 530, Ruedemann says : "It has been pointed out in a former publication that not only did there exist in the graptolites ontogenetic growth stages in the development of the individual zooids, .... but the rhabdosomes in toto and in their parts, the branches, seem also to pass through stages which suggest phylogcnetically preceding forms."

Of the various ways in which these astogenetic stages express them-

selves, Ruedemann mentions the following : "The original direction of growth of the branches of the Dichograptidss has been in the approximate

continuation of the sicula, i. e.. an ascending erect position as long as

the rhabdosomes were sessile, on the ground. These became pendant when the graptolites attached themselves in a suspended position to seaweeds,

as numerous hydroids do today. To restore to the zooids their original

. . . . erect position, the branches began now to recurve .... [becoming progressively horizontal, reflexed. reclined and recumbent]

. . . . We find now in the majority of the Dichograptiche with the

above cited growth directions of the branches, that the latter still retain

their original dependent direction, in the proximal parts in some species

. . . . while in others by the law of acceleration, the dependent proxi- mal direction has already changed into a horizontal one .... the change in direction becoming progressively more abrupt as the final direc- tion of the branches becomes reclined .... or recumbent The branches pass hence, in their development, through different directions

representing ontogenetic stages that repeat stations in their phylogenetic

development." (48.)

An analogous fact is found in the character of the thecte. "A com- parison of the form of the theese of the youngest dichograptid genera

. . . . with that of the older and presumably phylogenetically preced- 334 ing genera .... shows that in general the older genera have the more tubular, simpler thecal, with the less protected apertural margins. It is, hence, apparent that the stolonal or earlier thecse of the rhabdosomes represent indeed the older types of thecal form." (48.) Other Glasses.—The case of the larva of Antedon has already been referred to. As pointed out by Bather (1), the stem ossicles of the larval

Antedon are of a complex and specialized type, and in a general way re- semble the stem ossicles of the Bourgueticrinidse of the Upper Cretaceous.

It is held by Bather that the structures of the adult ancestors have been pushed back by acceleration to the larval stages of the existing Antedon.

Recapitulation is also shown in the anal plate of Antedon. The anal plate appears between two of the radials and on the same level with them.

Subsequently it is lifted out from between the radials, and the latter close beneath it. Still later the anal plate is resorbed entirely. That this is the recapitulation of an adult character and not of a larval character, as con- tended by Hurst, is shown by the fact that the oldest crinoids do not possess the anal plate at all. It appears from paleontological evidence that this plate first appeared above the level of the radials, that it gradually sank down between the two posterior radials, and that at a far later period

(at about the close of the Paleozoic) it gradually passed upward again as it does in Antedon, and eventually disappeared.

Jackson has shown that there is good evidence of recapitulation among the fossil echinoids (33). In most regions of the echinoid the develop- ment is obscured by the more or less extensive resorption, but the plates of the corona may show by their position and number, the course of devel- opment. Jackson holds that the introduction of columns of plates, both interambulacral, and ambulacral, in Melonites, etc., indicates the stages of

growth through which the individual has passed in its development. . He shows that two columns of ambulacral plates "may be accepted as the usual characteristic of the whole class, which finds its representative in the majority of the adults, in nearly all young, and in the adult of the simplest and oldest known type, Bothriocidaris." Interambulacral areas originate ventrally in a single plate. Only one genus is known, however, that has a single row of plates in the adult, namely Bothriocidaris. This is the simplest known and "perhaps the simplest conceivable echinoid." 335

In Goniocidaris the interambulacral plates of the adult are approx- imately hexagonal in form instead of pentagonal. "The relative form of the plates in young Goniocidaris is almost exactly the same as in the primitive type, Bothriocidaris."

"The early stage in which we find a single interambulacral plate, to- gether with two ambulacral plates, in each area is so important that it is desirable to give it a name, the proteehinus stage. The protechinus is an early stage in developing Echini, belonging to the phylembryonic period, in which the essential features of the echinoid structure are first evinced.

. . . . This protechinoid stage of Echinoderms is comparable as a stage in growth to a similar stage which is expressed in the protegulum of brachiopods, the protoconch of cephalous mollusks, the prodissoconch of pelecypods, and the protaspis of trilobites." (33.)

Miss Smith (Mrs. Alexander Shannon) has shown very conclusively the exact resemblance of the form of the young Pentremites conoideus to the adult Codaster (52). In Codaster the conical form, narrowest at the base and enlarging upward, is maintained throughout life. In Pentremites only the early stages of growth have this form, while the adult is broadest at the base and narrowest at the top.

This evidence from development would, according to the theory of recapitulation, indicate that Codaster stands in an ancestral relation to

Pentremites, and it is therefore of importance to the theory that Bather

(2) from other evidence has independently reached the same conclusion as Miss Smith in regard to the relationship of the two forms. 1

Among corals Beecher (5) has worked out the development of Pleu- rodictywm lenticulare and concludes that the first neanic stage, in the manner of growth and the structure of the corallum, is very suggestive of Aulopora, and should be given considerable significance." Girty (21) conies to the same conclusion from a study of Favosites forbesi, etc.

Bernard (14) has shown that the coral colony in similar fashion to the bryozoan colony and the graptolite colony behaves as an individual.

In another paper (13) he has recognized as the first growth stage of the

1 Bather's conclusion was published in 1900, and Miss Smith's paper in 1906. The latter, however, was not aware of Bather's views as to the relationships of these two forms, so that the conclusions of the two workers, arrived at indepen- dently and from different lines of evidence are all the more important and convinc- ing. Bather says in a review of Miss Smith's paper that he considers Pentremites as the "extreme link in the series Codaster—Pliaenoschisma—Cryptoschisma—Oroph- ocrinus—Pentremitidea—Pentremites." 336

coral skeleton the "prototheka," or basal cup of the first individual of the colony. x

Lang (38) has written a very suggestive paper on growth stages of British species of corals, in which he points out the fact that the onto- genetic stages are repeated in each rejuvenescence (branching?), and sug- gests that we have here an example of localized stages in development

(see Jackson 34). It may be remarked at this point that Ruedemann has

also detected localized stages in graptolites (47, 48), and Lang in Bryozoa

(36). Lang also, in the paper on corals, concludes that there is recapitu- lation in the coral genera studied by him. of ancestral characters, and he

gives a table illustrating this. 2

Summary.—Paleontologists almost universally accept the theory of re- capitulation. Its chief critics have been embryologists. The reason for

the difference in attitude is probably to be sought in the fact that the former ordinarily compare epembryonic stages with adult characters of

geologically older species, while the latter too often compare embryonic

stages with the adult stages of existing species. It is also to be noted

that in recapitulation we have to do with morphological and not with

physiological characters, and that the row of cells from the egg to the adult may be morphologically the same in two organisms, while being at the same time physiologically different. Until it can be shown that two organ-

isms morphologically different in the adult must of necessity be morpho- logically different at all stages, the argument of Montgomery, Hurst and others proves nothing.

1 The term prototheka was proposed simultaneously (January, 1904) by Ber- nard and myself for the earliest skeletal structure of. the coral colony. We have used it, however, in a slightly different sense. Bernard applies it not only to the first individual of the colony, but also to the basal plates or cups of later indi- viduals. I intended to restrict it to the basal cup of the first individual. The references are as follows : Bernard, H. M., The prototheka of the Madreporaria, with special reference to the genera Calostylis, Linds., and Mosleya, Queleh. Ann. Mag. Nat. Hist., Ser. 7, vol. xiii, Jan. 1904. Cumings, E. R., The development of some Paleozoic Bryozoa. Am. Jour. Sci., vol. xvii, Jan., 1904 (footnote, p. 74). 2 This so-called rejuvenescence in corals appears to be a species of budding, in which the bud is directly superimposed upon the parent. It is fission occurring in a horizontal plane, as suggested by Bernard (14), and the new skeleton is in direct continuity with the old. This is the same idea exactly as that advanced by Ulrich some years ago (63) to account for the diaphragms of the Bryozoa Trepostomata. In the case of the Trepostomata the zoceeium is frequently oper- culate (ex. Callopora). and there is good evidence that the bud grows up through the operculum hence leaving it behind as the floor of the new individual. 337

In the Cephalopoda, Peleeypoda, Gastropoda, Braehiopoda, Trilobita, Bryozoa, Graptolites. Echinoderuis aud Corals, examples are pointed out in which there is clear and unmistakable evidence of recapitulation. In most of these cases it is the epembryonic and not the embryonic stages that are the basis of comparison. Paleontological Laboratory, Indiana University, Bloomington, Indiana.

REFERENCES.

1. Bather, F. A.—The recapitulation theory in Paleontology. Nat. Sci..

ii. 1893.

2. Treatise on Zoology. Echinoderma, 1900.

3. Barrande. J. — Sao Hirsute, Barrande, ein Bruchstiick aus dein "Sys-

teme Silurien du centre de la Boheme." Neues Jahrb. f. Min. Geol. u. Pal., 1849.

4. Systeme Silurian du centre de la Boheme," 1-iere partie. 1852.

5. Beecher, C. E.—Development of a Paleozoic poriferous coral. Trans. Conn. Acad. Sci., vol. viii, 1891.

6. Development of the Braehiopoda. Part I, Introduction. Am. Jour. Sci., (3), xli, 1891; Part, II, Classifications of the stages of growth and decline. Am. Jour. Sci., (3), xliv, 1892; Part III, Morphology of the Brachia. Bull. U. S. Geol. Surv.. No. 97, ch. iv, 1897.

7. Some correlations of ontogeny and phylogeny in the Braehiopoda. Am. Nat, xxvii. 1893.

8. Larval stages of Trilobites. Am. Geol.. vol. xvi, 1895.

9. Outline of a natural classification of the trilobites. Am. Jour.

Sci., (4), iii. 1897.

10. Studies in Evolution. (Mainly reprints of occasional papers. The papers listed above are also contained in this collection.) New York, Charles Scribner's Sons. 1901.

11. and Clarke. J. M.—The development of some Silurian Braehi-

opoda. Mem. N. Y. State Mus.. No. i. 1889. 12. —— and Schuchert, C.—Development of the brachial supports in Die-

lasma and Zygospira. Proc. Biol. Soe.. Washington, vol.. viii. 1893.

13. Bernard. H. M.—The prototheka of the Madreporaria. with special reference to the genera Colostylis, Linds. and Mosleya, Qneleh.

Ann. Mag. Nat. Hist, Ser. 7, vol. xiii, 1904.

[22-23003] 338

14. British Museum Catalogue of Madreporarian corals, vol. vi, 1906.

15. Cope, E. D.—Origin of the fittest. New York, D. Appleton & Co., 1887. 16. Cannings, E. R.—The morphogenesis of Platystrophia. A study of the evolution of a Paleozoic brachiopod. Am. Jour. Sci., (4), vol. xv, 1903.

17. Development of some Paleozoic Bryozoa. Am. Jour. Sci., (4), vol. xvii, 1904.

18. Development of Fenestella. Am. Jour. Sci., (4), vol. xx, 1905.

19. Dall. W. H.—Trans. Wagner Free Inst., vol. iii. 1S90, p. 67. (Quoted by Grabau.) 20. Eigenmann, C. H.—The eyes of the blind vertebrates of North Amer- ica. Mark Anniversary Volume, Article ix, 1903. 21. Girty, G. H.—Development of the corallum in Favosites forbesi var. occidentalis. Am. Geol., vol. xv, 1S95. 22. Grabau, A. W —Studies of Gastropoda. Am. Nat., xxxvi. 1902.

23. Studies of Gastropoda II. Fulgur and Sycotypus. Am. Nat., vol. xxxvii, 1903.

24. Phylogeny of Fusus and its allies. Smithsonian Misc. Coll., vol. xliv, 1904.

25. Studies of Gastropoda III. On orthogenetic variation. Am. Nat, vol. xli. 1907. 26. and Shinier.— See Shinier and Grabau. 27. Greene, F. C.—The development of a Carboniferous brachiopod, Cho- netes granulifer, Owen. Jour. Geol., vol. xvi, 1908.

28. Griggs, Pt. F.—Juvenile Kelps. Am. Nat., Jan. and Feb., 1909. 29. Holm, G.—On Didymograptus Tetragraptus and Phyllograptus. Geol.

Mag., Dec. iv. vol. ii, 1S95.

30. Hurst, —The recapitulation theory. Nat. Sci., ii. 1893. 31. Hyatt, A.—The phylogeny of an acquired characteristic. Proc. Am. Phil. Soc. vol. xxxii (Anniversary volume). 1894. (See other references to this author in the footnote at the close of the section on the Cephalopoda.) 32. Jackson. R. T.—Phylogeny of the Pelecypoda. the Avieulida? and their allies. Mem. Bos. Soc. Nat. Hist., vol. iv. 1S90.

33. Studies of Paleehinoidea. Bull. Geol. Soc. Am., vol. vii. 1896.

34. Localized stages in development in plants and animals. Mem. Bos. Soc. Nat. Hist., vol. v. No. 4. 1S99.

35. Lang. W. D.—The Jurassic forms of the 'genera.' Stomatopora and

Proboscina. Geol. Mas:.. Dec. v. vol. i, 1904. 339

30. Stomatopora antiqua, Haime and its related forms. Geol. Mag.,

Dec. v, vol. ii, 1905.

37. The reptant Eleid Polyzoa. Geol. Mag., Dec. v, vol. iii, 1900.

38. Growth stages in the British species of the coral Parasniilia. Proc. Zool. Soc, London, 1909.

39. Matthew, G. F.—Illustrations of the fauna of the St. John group con-

tinued : On Conocoryphea with further remarks on Para-

doxides. Trans. Roy. Soc, Canada, vol. ii, sec. iv, 1884.

40. Illustrations of the fauna of the St. John group, No. 4, pt. ii, the smaller Trilobites with eyes (Ptychoparidae and Ellipso-

cephalidse). Trans. Roy. Soc. Canada, vol. v, sec. iv, 1S87.

41. Sur le Development des premiers Trilobites. Ann. Soc. Malac. de Belgique, 1889. 42. Montgomery, T. C.—An analysis of racial descent. 1900. 43. Morgan. T. H—Evolution and adaptation. 1903. 44. Nitsche, H.—Beitrage zur Kentniss der Bryozoen. Zeitschr. fur Wiss. Zool., Bd. xxi, 1871.

45. Pergens, E.—Untersuchungen an Seebryozoen. Zool. Anzeig., Jg. xii, 1889. 40. Raymond, P. E. —Developmental change in some common Devonian brachiopcds. Am. Jour. Sci., vol. xvii, 1904. 47. Ruedemann, R.—Growth and development of Goniograptus thureaui, McCoy. Bull. N. Y. State Mus., No. 52, 1902.

48. Graptolites of New York, Pt. I. Mem. N. Y. State Mus., No. 7, 1904. 49. — Graptolites of New York, Pt II. Mem. N. Y. State Mus., No. 11, 1908. 50. Sardeson, F. W.—The phylogenic stage of the Cambrian Gastropoda. Jour. Geol., vol. xi, 1903. 51. Shinier, H. W., and Grabau, A. W.—Hamilton group of Thedford, Ontario. Bull. Geol. Soc. Amer., vol. xiii, 1902. 52. Smith, Essie (Mrs. Alexander Shannon).—The development and vari- ation of Pentremites conoideus. Indiana Dept. Geol. Nat. Res., 30th Ann. Rept., 1900. 53. Smith, Burnett. —Phylogeny of the species of Fulgur with l'emarks on an abnormal specimen of Fulgur canaliculatum, and sexual dimorphism in Fulgur carica. Proc. Acad. Nat. Sci., Phila.,

vol. liv, 1902.

54. Pbylogeny of the races of Yolutilithes petrosus. Proc. Acad. Nat. Sci., Phila.. vol. Iviii. 1900. 340

55. New species of Athleta and a note on the morphology of Athleta petrosa. Proc. Acad. Nat. Sci. Phila., vol. lix, 1907.

56. A contribution to the morphology of Pyrula. Proc. Acad. Nat. Sci. Phila., vol. lix, 1907.

57. Smith. J. P.—The biogenetic law from the standpoint of Paleontology. Jour. Geol., vol. viii, 1900.

58. The development and phylogeny of Placenticeras. Proc. Cali-

fornia Acad. Sci., (3), vol. i, No. 7, 1900. (For references to other papers by J. P. Smith, see the footnote at the close of the section on Gastropoda.) 59. Walcott, C. D.—Fossils of the Utica slate, and metamorphoses of Triarthrus becki. Trans. Alb. Inst., vol. x. 1S79.

60. Second contribution to the studies of the Cambrian faunas of North America. Bull. U. S. Geol. Snrv., No. 30, 18S6.

61. The fauna of the Lower Cambrian or Olenellus zone. 10th Ann. Rept. U. S. Geol. Surv., 1888-9.

62. Cambrian geology and paleontology. No. 4, Classification and terminology of the Cambrian Brachiopoda. Smithsonian

Misc. Coll.. vol. liii, 1908.

63. Ulrich, E. O—Bryozoa. Geol. Surv. 111., vol. viii, 1890.

64. Zeleny, C.—Compensatory regulation. Jour. Exper. Zool., vol. ii. No. 1, 1905.

65. Zittel, K.—Paleontology and the biogenetic law. Nat. Sci., vol. vi, 1895.